Struct FrlTrainer 

pub struct FrlTrainer<C, P> { /* private fields */ }

The central async training type. Owns an async ScdcClient and an async HdmiPhy.

The async counterpart to plumbob::FrlTrainer: it runs plumbob’s state machine, plumbob::lts::run, awaiting each SCDC and PHY operation, so it yields to the executor wherever the I/O waits. Reusable across training attempts; use into_parts to recover the SCDC client and PHY.

Cancellation

train is not cancel-safe. Dropping its future — a select! against a hot-unplug, a timeout — stops training wherever it is: the sink may be configured for an FRL rate and the PHY driving training patterns, and no LTS:L runs, since a future cannot await anything once dropped. Call exit_to_tmds afterwards to return both ends to TMDS. Bounding training is better left to the poll limits in TrainingConfig, which end every attempt in a defined state.

Implementations

impl<C: ScdcClient, P: HdmiPhy> FrlTrainer<C, P>

pub fn new(scdc: C, phy: P) -> Self

Constructs a new FrlTrainer owning the given async SCDC client and PHY.

pub fn into_parts(self) -> (C, P)

Consumes the trainer and returns the SCDC client and PHY.

pub fn scdc(&self) -> &C

The SCDC client.

pub fn scdc_mut(&mut self) -> &mut C

The SCDC client, mutably.

pub fn phy(&self) -> &P

The PHY.

pub fn phy_mut(&mut self) -> &mut P

The PHY, mutably: after Success, for example, to start video with set_frl_output(FrlOutput::Active).

pub async fn exit_to_tmds( &mut self, ) -> Result<(), ExitError<C::Error, P::Error>>

LTS:L on demand: stops the training patterns, returns the PHY to TMDS, turns FRL off in Config_1 and clears FLT_update if it is set.

The async counterpart to plumbob::FrlTrainer::exit_to_tmds: for taking an FRL link down when the display is disabled or unplugged, before a mode change, after an error that did not lead to LTS:L (TmdsExit::Skipped or HeldUnderTest), and after a dropped train future (see Cancellation). Safe to call in any state, and again. Every step is attempted even when one fails.

pub async fn train_at_rate( &mut self, rate: HdmiForumFrl, config: &TrainingConfig, ) -> Result<Trained, TrainingError<<C as ScdcClient>::Error, <P as HdmiPhy>::Error>>

Trains at a single rate: train(&[rate], config).

pub async fn train( &mut self, rates: &[HdmiForumFrl], config: &TrainingConfig, ) -> Result<Trained, TrainingError<<C as ScdcClient>::Error, <P as HdmiPhy>::Error>>

Trains over rates in order, stepping down when the sink requests it.

The async counterpart to plumbob::FrlTrainer::train, with the same outcomes and errors. Not cancel-safe: see Cancellation.

pub async fn train_with_events<F: FnMut(TrainingEvent)>( &mut self, rates: &[HdmiForumFrl], config: &TrainingConfig, record: &mut F, ) -> Result<Trained, TrainingError<<C as ScdcClient>::Error, <P as HdmiPhy>::Error>>

Like train, calling record with each TrainingEvent as it occurs: the events train_traced (with alloc) collects, without an allocator. record is a plain callback; it is called between I/O operations, never while one is awaited.